→ See a full sample report first: Free sample Reviewer Zero — Local Edition · the reviewer that reads your paper before Reviewer 2 does The complete methodology audit engine as installable skills. Runs locally in the AI you already use - Claude Code, Cursor, Codex, Deepseek, Kimi, and more. One-timeYour reviewer will ask whether you tested measurement invariance. Whether that control variable is a confounder or a mediator. Whether "leads to" is earned by a cross-sectional design. It asks first — in whichever AI you already use — so you fix it before it's a rejection.You get back a structured audit: every finding rated Fatal, Major, or Minor; a gap list keyed to your journal's own checklist (CONSORT, STROBE, COREQ, GRAMMS…); and a limitations paragraph you can drop into the manuscript.A glimpse of the outputPaste a Methods section, get back something like:🔴 Fatal — "Psychological safety leads to innovative behavior" isn't supported by a single-wave cross-sectional survey. Reframe as an association, or add a second wave.🟠 Major — "Leader support" is controlled for, but it's a mediator, not a confounder.🟢 Minor — Report McDonald's ω, not α alone.…plus a STROBE gap list (by item number) and a ready-to-paste limitations paragraph.Works whether you've written it or notHave a draft? Paste your Methods section (or a proposal, or a pre-registration) and get the full report in about two minutes. Ideal for a resubmission or a deadline.Still designing the study? Start from your research question. Reviewer Zero walks you through the design one step at a time, beginning at conceptualization, and helps you fix problems before you've collected a single data point. The cheapest stage to catch a fatal flaw.One for each kind of studyQuantitative — construct validity through DAGs and identification (IV/DID/RDD), measurement invariance, McDonald's ω over α, HTMT over Fornell-Larcker.Qualitative — paradigm coherence, information power (not "30 for grounded theory"), reflexivity, Tracy's big-tent criteria, and the reflexive-vs-coding-reliability trap (demanding κ from a study that isn't designed to produce it) that gets papers desk-rejected.Mixed-methods — the only one that tests whether your integration is real, or whether you wrote two parallel studies and stapled them together. Drafts your joint display.Case-based & set-theoretic — process tracing (the four evidential tests: hoop, smoking gun, straw-in-the-wind, doubly decisive; mechanism specification; DA-RT transparency) and QCA (calibration, necessity vs sufficiency, consistency/PRI/coverage, logical remainders).Pure theory - formal and game-theoretic models, conceptual frameworks, critical essays. Audits assumptions (which are load-bearing?), derivation soundness, the tautology trap, scope conditions, and falsifiability judged against the model's declared purpose.A router picks the right one if you're not sure.What makes it different from asking ChatGPT to review your methodsEvery threshold cites its source. α > .7, r > .6, κ > .7 are conventions, not laws — Reviewer Zero tells you which, with the citation, so you can defend a number to a reviewer instead of parroting it.It maps to the checklist your journal actually mandates. Reviewers audit against STROBE or COREQ, not a textbook. Every report ends with a gap list keyed to item numbers.It rates severity. Not a wall of 30 nitpicks — a ranked list of which one sinks the paper and which three are one-sentence fixes.It gets the contested cases right. Likert-as-interval, saturation vs information power, face validity's real status, "control for everything" as a mistake. This is where generic AI advice is confidently wrong.What's in the box 5 method skills plus 1 router (install in Claude Code or Cursor, or paste as a system prompt anywhere) Self-contained system prompts for DeepSeek, Kimi, GLM, Qwen, ChatGPT, and any OpenAI-compatible API 6 worked examples: 3 fast-mode audits of realistic manuscripts plus 1 coach-mode walk-through of designing a study from scratch A 126-point printable checklist and a visual diagnostic flowchart (light and dark) Plain-English license and an install guide for every platform Runs onClaude Code · Claude Desktop · Cursor · Codex · ChatGPT · DeepSeek · Kimi · Zhipu GLM · Qwen · MiniMax · any OpenAI-compatible API.Honest about what it isReviewer Zero is a methods coach, not peer review and not a guarantee. It runs on an LLM, so it can miss things and raise false alarms — treat every finding as a prompt for your judgment. It's built to help you think about your study, not to write it for you: the drafts it produces are scaffolding to rewrite in your own voice, and you should disclose AI use as your journal requires.One-time purchase ($199). Updates included. → Prefer to start with a subscription? Try the web version from $5/months: https://zenith6100.gumroad.com/l/cetey